Cloning, Expression, and Purification of the K5 Capsular Polysaccharide Lyase (KflA) from Coliphage K5A: Evidence for Two Distinct K5 Lyase Enzymes

Abstract

The Escherichia coli K5 capsular polysaccharide [-4)-βGlcA-(1,4)-αGlcNAc-(1-] is a receptor for the capsule-specific bacteriophage K5A. Associated with the structure of bacteriophage K5A is a polysaccharide lyase which degrades the K5 capsule to expose the underlying bacterial cell surface. The bacteriophage K5A lyase gene (kflA) was cloned and sequenced. The kflA gene encodes a polypeptide with a predicted molecular mass of 66.9 kDa and which exhibits amino acid homology with ElmA, a K5 polysaccharide lyase encoded on the chromosome of E. coli SEBR 3282. There was only limited nucleotide homology between the kflA and elmA genes, suggesting that these two genes are distinct and either have been derived from separate progenitors or have diverged from a common progenitor for a considerable length of time. Southern blot analysis revealed that kflA was not present on the chromosome of the E. coli strains examined. In contrast, elmA was present in a subset of E. coli strains. Homology was observed between DNA flanking the kflA gene of bacteriophage K5A and DNA flanking a small open reading frame (ORF_L) located 5′ of the endosialidase gene of the E. coli K1 capsule-specific bacteriophage K1E. The DNA homology between these noncoding sequences indicated that bacteriophages K5A and K1E were related. The deduced polypeptide sequence of ORF_L in bacteriophage K1E exhibited homology to the N terminus of KflA from bacteriophage K5A, suggesting that ORF_L is a truncated remnant of KflA. The presence of this truncated kflA gene implies that bacteriophage K1E has evolved from bacteriophage K5A by acquisition of the endosialidase gene and subsequent loss of functional kflA. A (His)₆-KflA fusion protein was overexpressed in E. coli and purified to homogeneity with a yield of 4.8 mg per liter of bacterial culture. The recombinant enzyme was active over a broad pH range and NaCl concentration and was capable of degrading K5 polysaccharide into a low-molecular-weight product.

At least 80 distinct capsular polysaccharides (K antigens) have been described previously for Escherichia coli (26). Expression of a polysaccharide capsule on the cell surface confers resistance to host immune defenses and other adverse environmental conditions (3, 29). In addition, many of these structurally distinct capsules act as receptors for bacteriophages, and variation of capsular structure may be a mechanism for evasion of bacteriophage infection. Due to their specificity, capsule-specific bacteriophages can be used for identification of numerous E. coli K serotypes. Integral to the tail structure of many capsule-specific bacteriophages are enzymes which degrade the bacterial polysaccharide and provide the bacteriophages access to receptors on the cell surface (5, 32–35).

The E. coli K5 capsular polysaccharide is a polymer of -4)-βGlcA-(1,4)-αGlcNAc-(1- (40). This structure is identical to N-acetylheparosan, the precursor polymer of heparin and heparan sulfate (17). Coliphage K5A, a bacteriophage specific for E. coli K5 capsule-expressing strains, has been previously described (9). This bacteriophage contains a lyase which degrades the K5 polysaccharide randomly throughout the polymer by a β elimination reaction. The final reaction products of this bacteriophage lyase consist of hexa-, octa-, and decasaccharides (8, 12). In addition to the bacteriophage-borne K5 lyase, a chromosomally encoded K5 lyase enzyme has been described for E. coli SEBR 3282 (16). The chromosomal gene (elmA) has been cloned and expressed in E. coli K-12 and has been shown to produce final reaction products with higher molecular weights than those of products observed for the bacteriophage lyase (12, 16).

In this communication, we report the cloning and analysis of the bacteriophage K5A lyase gene and describe the expression, purification, and characterization of the recombinant enzyme. We show that the bacteriophage K5A lyase gene is not found on the chromosome of E. coli K5 strains, whereas elmA is present on the chromosome of some strains. The difference in the distribution of the bacteriophage K5A lyase gene and elmA and lack of any shared nucleotide homology indicate that these two genes are distinct variants. These genes either have diverged extensively throughout the evolution of E. coli or have convergently evolved from separate ancestral genes. Additionally, we provide evidence indicating that bacteriophage K5A is likely to be the progenitor to the E. coli K1 capsule-specific bacteriophage K1E.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

E. coli K-12 SURE {e14⁻ mcrA Δ(mcrCB-hsdSMR-mrr)171 endA1 supE44 thi-1 gyrA96 relA1 lac recB recJ sbcC umuC::Tn5 uvrC [F′ proAB lacI^q ZΔM15 Tn10]} was obtained from Stratagene, and E. coli K-12 TOP10 ([F⁻ mcrA Δ(mcrCB-hsdSMR-mrr) φ80lacZΔM15 ΔlacX74 deoR recA1 araD139 Δ(araA-leu)7697 galU galK rpsL endA1 nupG]) was obtained from Invitrogen. E. coli K5 wild-type strains were obtained from K. Jann, Max Planck Institut für Immunbiologie, Freiburg, Germany. Plasmids used in this study were pTTQ18 (31) and pBAD/HisB (Invitrogen). Bacteria were routinely grown at 37°C in Luria-Bertani (LB) medium supplemented with 100 μg of ampicillin per ml when required. Bacteriophage K5A was propagated using E. coli Bi8337-41 as a host in LB medium containing 0.2% d-glucose and 10 mM MgCl₂.

DNA methods.

Bacteriophage DNA was purified by the procedure described for bacteriophage lambda (30). Recombinant DNA techniques were performed according to standard procedures (30). Restriction and modifying enzymes were purchased from Boehringer, Mannheim, Germany. Double-stranded DNA sequencing was performed with custom oligonucleotide primers using the ABI PRISM BigDye Terminator cycle sequencing kit together with an Applied Biosystems automated DNA sequencer.

To construct the (His)₆-KflA fusion protein, the kflA gene was amplified from pBL4 by PCR using Pfu DNA polymerase. Two primers, 5′AGGAAGATCTATGGCTAAATTAACCAAACCT3′ and 5′TGCCGAATTCTTACTTAGGAAGGGCAGCTAG3′, were constructed incorporating BglII and EcoRI restriction sites into the 5′ and 3′ ends of the kflA gene, respectively. The amplification product was ligated into BglII-EcoRI-digested pBAD/HisB to position kflA in frame with the N-terminal (His)₆ fusion tag of the vector.

The nucleotide sequences of the kflA and elmA genes were compared using the BLAST 2 program, which aligns two given nucleotide sequences (37).

Small-scale recombinant protein expression and preparation of cell lysates.

The bacteriophage genomic library was screened for lyase activity by assaying lysates from pools of 10 recombinant clones in E. coli SURE. Individual clones were then screened from the pools showing lyase activity. E. coli SURE cells (10 ml) containing the genomic fragments in pTTQ18 were grown to an optical density at 600 nm (OD₆₀₀) of 0.6 at 37°C with shaking at 200 rpm and then induced by addition of 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). Incubation was then continued for 1 h, and the cells were harvested at 5,000 × g for 10 min. The cells were washed once in ice-cold 100 mM Tris-Cl (pH 7.5), resuspended in a 1/10 volume of the same buffer, and then sonicated with five 30-s bursts separated by 1 min of cooling in ice water. The sonicates were clarified by centrifugation at 10,000 × g for 20 min.

Arabinose-dependent expression of the (His)₆-KflA fusion protein was performed using E. coli TOP10 as an expression host. This strain is able to import l-arabinose but unable to metabolize this sugar. For small-scale expression of fusion protein TOP10(pLYA100), cultures were grown at 37°C at 200 rpm in 10 ml of LB medium to an OD₆₀₀ of 0.5. l-Arabinose was then added, and the cultures were incubated for a further 2 h. One milliliter of each culture was harvested, and the pellets were resuspended in lysis buffer (100 μl of 125 mM Tris-Cl, 0.5% [wt/vol] sodium dodecyl sulfate [SDS], 5% [vol/vol] glycerol, 1.25% [vol/vol] β-mercaptoethanol, 0.05% [wt/vol] bromophenol blue, pH 6.8) and boiled for 5 min. These lysates (10 μl) were used directly for SDS-polyacrylamide gel electrophoresis (PAGE).

Purification of K5 polysaccharide.

Bi8337-41 cells were grown overnight in 500 ml of LB broth and harvested at 5,000 × g for 10 min. The pellet was washed once in 50 ml of phosphate-buffered saline (pH 7.2), resuspended in 50 ml of extraction buffer (50 mM Tris-Cl, 5 mM EDTA, pH 7.3), and incubated at 37°C for 30 min. The cells were pelleted by centrifugation and treated three times further with extraction buffer. Polysaccharide was precipitated from the pooled supernatants by addition of cetyl-3-ethyl ammonium bromide (Na salt) to a final concentration of 0.1% (wt/vol) followed by incubation at room temperature for 16 h. The precipitate was recovered by centrifugation at 10,000 × g for 20 min at 20°C, dissolved in 1 M NaCl, and precipitated again by addition of ethanol to 80% (vol/vol). The precipitate was dissolved in 5 ml of distilled water and dialyzed against distilled water. The solution was centrifuged at 100,000 × g, and the supernatant containing K5 polysaccharide was freeze-dried.

K5 polysaccharide lyase assays.

Polyacrylamide gel assays were performed as previously described (27). Briefly, lysate or purified fusion protein was incubated in 25 μl of 25 mM Tris-acetate (pH 8.5) containing 20 μg of K5 polysaccharide for 1 h at 37°C. A 1/10 volume of 2 M sucrose–0.02% (wt/vol) bromophenol blue in electrophoresis buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA, pH 8.3) was added, and 20 μl of each reaction mixture was loaded onto a polyacrylamide gel (25% acrylamide, 0.82% bisacrylamide). Gels were pre-electrophoresed for 1 h at 12 V/cm, and samples were electrophoresed at 6 V/cm. Gels were stained by the combined alcian blue-silver staining method (24).

The spectrophotometric assay for lyase activity took advantage of the Δ^4,5 bond formed due to cleavage of the K5 polysaccharide by β elimination (2, 12). The A₂₃₂ was monitored as a function of time using saturating concentrations of K5 polysaccharide as substrate. The reaction rate was expressed in absorbance units (AU) per minute. Typical reactions were performed at 37°C and contained 1 μg of purified KflA fusion protein and 800 μg of K5 polysaccharide in a 1-ml reaction mixture. The rate of product formation was monitored over 5 min. Buffer conditions were varied according to the nature of the specific experiment.

Purification of the (His)₆-KflA fusion protein.

E. coli TOP10(pLYA100) (500 ml) was grown in LB broth supplemented with ampicillin at 37°C at 200 rpm. When the OD₆₀₀ reached 0.5, l-arabinose was added to a concentration of 0.02% (wt/vol) and the culture was incubated for a further 2 h. The bacteria were harvested at 4,000 × g for 10 min at 4°C. The pellet was washed in 200 ml of buffer A (50 mM Tris-Cl, pH 8.5) and resuspended in 40 ml of ice-cold buffer A. The cell suspension was passed once through a French pressure cell operated at 20,000 lb/in². The resulting lysate was then centrifuged at 100,000 × g for 1 h at 10°C.

The entire lysate was loaded on a 5-ml Pharmacia HiTrap Q column (pre-equilibrated in buffer A) at 2 ml/min. This was followed by washing with 100 ml of buffer A at 5 ml/min. Elution was performed over 150 ml with a linear 0 to 0.5 M NaCl gradient (in buffer A). Fractions (2 ml) were collected, and those containing the most fusion protein, as determined by SDS-PAGE and Western blotting with α-Xpress antibody, were pooled.

The pooled solution from the anion-exchange chromatography was adjusted to 500 mM NaCl and 0.1% (vol/vol) Triton X-100. This was then loaded at 3 ml/min onto a 5-ml Pharmacia HiTrap chelating column (charged with Ni²⁺), and the column was washed with 100 ml of buffer B (50 mM Tris-Cl, 500 mM NaCl, 0.1% [vol/vol] Triton X-100, pH 8.5). Proteins were eluted into 2-ml fractions at 5 ml/min with a linear 0 to 500 mM imidazole gradient (in buffer B) over 150 ml. Fractions judged to contain the KflA fusion protein by SDS-PAGE and Western blotting were pooled and concentrated in a Centriprep 30 concentrator (Amicon) according to the manufacturer's instructions. The concentrated eluant was desalted in buffer C (50 mM Tris-Cl, 0.1% [vol/vol] Triton X-100, pH 8.0) using four 5-ml Pharmacia HiTrap desalting columns linked in a series. The desalted eluant was concentrated as described above to 5 ml and stored at 4°C.

Additional analytical methods.

SDS-PAGE was performed according to the method of Laemmli (13), and Western blotting was performed as described by Towbin et al. (38) using the ECL detection kit (Amersham). Protein concentration was estimated with the Bio-Rad protein assay kit using bovine gammaglobulin as a protein standard.

Nucleotide sequence accession number.

The kflA nucleotide sequence has been submitted to the GenBank database under the accession no. Y10025.

RESULTS AND DISCUSSION

Cloning and nucleotide sequencing of the coliphage K5 lyase gene.

A library of bacteriophage K5A DNA was constructed in the expression vector pTTQ18 and transformed into E. coli SURE. Lysates from 250 recombinants were screened for lyase activity by overnight incubation with purified K5 polysaccharide followed by PAGE of the reaction products. One clone, pBL4, exhibited detectable lyase activity (data not shown). Plasmid pBL4 contained a cloned fragment of 2.8 kb, and Southern blot analysis confirmed that the cloned DNA was derived from the genome of bacteriophage K5A (data not shown).

The nucleotide sequence of the cloned DNA in pBL4 was determined. The insert comprised 2,872 bp with a G+C content of 48%. Two open reading frames (ORFs) were identified in the opposite orientation to the tac promoter of pTTQ18 (Fig. 1). The largest ORF (ORF₆₃₅) was 1,905 bp and encoded a putative protein of 635 amino acids with a predicted molecular mass of 66.9 kDa. A putative ribosomal binding site was identified 10 bp upstream of the ORF₆₃₅ translational initiation codon. A BLAST search (1) of the GenBank peptide database revealed 53% identity over 619 amino acids (68% similarity) between the ORF₆₃₅ polypeptide and ElmA, a chromosomally encoded K5 polysaccharide lyase from E. coli SEBR 3282 (O10:K5:H4) (data not shown) (16). This amino acid homology indicates that ORF₆₃₅ encodes the bacteriophage K5A lyase, and therefore, this gene was designated kflA (K5 lyase). Located 3′ to kflA was a partial ORF (ORF_P) encoding 147 amino acids (Fig. 1). ORF_P was also preceded by a putative ribosomal binding site. A BLAST search of the GenBank database failed to find any proteins with significant homology to the polypeptide encoded by ORF_P. The noncoding region of the cloned bacteriophage K5A DNA 5′ to kflA contained homology to the consensus for SP6 bacteriophage promoters as described previously for the K1E bacteriophage (6, 22, 23). This implies that kflA is transcribed by a bacteriophage RNA polymerase.

FIG. 1.

Comparison between the cloned coliphage K5 DNA sequence and that of bacteriophage K1E encoding the endosialidase gene. (A) Arrows indicate ORFs and the direction of translation. ORF_P is translated in the same direction as the K5 polysaccharide lyase gene (kflA). Regions of nucleotide homology between the two bacteriophage sequences are depicted by similarly shaded boxes joined by dotted lines. (B) A BLAST comparison showing homology between the N termini of the deduced polypeptides encoded by ORF_L and kflA. Conservative amino acid changes are represented by “+.”

The kflA gene is distinct from elmA.

Comparison of the nucleotide sequences of kflA and elmA using a pairwise BLAST2 alignment (36) identified a short stretch of homology of 77% over 69 bases (data not shown). The lack of extensive homology suggests that these two genes are distinct and either have been derived from separate progenitors or have diverged from a common progenitor for a considerable length of time. To determine if a chromosomal variant of kflA exists, E. coli K5 wild-type isolates were analyzed by Southern blot analysis using the kflA gene as a probe. No hybridization was detected at low stringency among the strains examined (data not shown). Therefore, it is apparent that KflA is not encoded on the chromosome of E. coli strains and has evolved solely as a phage-encoded enzyme distinct from elmA. Southern hybridization of genomic DNA from the E. coli K5 wild-type strains using the elmA gene as a probe at high stringency revealed hybridization to a fragment of approximately 11.5 kb in 4 out of 11 of these strains (Fig. 2). Therefore, elmA is not present on the chromosome of all E. coli K5 strains. Three of the E. coli strains hybridized by the elmA probe (20026, 21786, and 21834) (Fig. 2, lanes 2, 5, and 9) have been shown to contain K5 lyase activity in cell lysates and, in addition, produce shorter polymer on their surfaces than those of other K5 strains (11). It is not known whether the O15:K5 strain, which also hybridized to elmA (Fig. 2, lane 8), exhibits intracellular lyase activity or low-molecular-weight polysaccharide. The correlation between K5 polymer size and the presence of elmA suggests that the chromosomally encoded lyase may have a regulatory role in capsule synthesis. The relationship between the presence of elmA in E. coli and expression of low-molecular-weight K5 capsule with the ecology and pathogenicity of such strains is unclear. It is possible that ElmA may be encoded by a cryptic K5-specific bacteriophage (distinct from K5A) which is present only in a subset of K5 strains. A role for a chromosomally encoded ElmA lyase in control of capsule size would be different from the role of the KflA enzyme, which is utilized solely by bacteriophage K5A for recognition of E. coli cells expressing K5 capsule and subsequently to gain access to the cell surface.

FIG. 2.

Southern blot showing hybridization of the elmA gene with HindIII-digested genomic DNA of various E. coli wild-type strains. Lane 1, Bi8337-41 (O10:K5:H4); lane 2, 20026 (O10:K5:H4); lane 3, 21831 (O86:K5:H10); lane 4, 21835 (O1:K5:H1); lane 5, 21786 (O4:K5); lane 6, 21795 (O5:K5:H4); lane 7, 21195 (O6:K5); lane 8, 21832 (O15:K5); lane 9, 21834 (O117:K5); lane 10, 21830 (O25:K5:H1); lane 11, 21836 (O75:K5). Strain designations refer to the Freiburg collection number. The elmA probe was amplified by PCR from the chromosome of strain 20026. The size of the restriction fragments that hybridized to the probe is indicated on the right.

Coliphage K5 is the progenitor of bacteriophage K1E.

Analysis of the 5′ noncoding region of the cloned bacteriophage K5A nucleotide sequence revealed that the first 437 nucleotides, up to the kflA translation initiation codon, exhibited 88% identity with sequence 5′ to a small ORF (ORF_L) encoded by the E. coli K1 capsule-specific bacteriophage K1E (Fig. 1A) (6, 22). In bacteriophage K1E, ORF_L is located immediately 5′ to an ORF encoding the endosialidase which degrades the polysialic acid capsule of E. coli K1 (22). In addition, the nucleotide sequence spanning 88 bp between the 3′ end of kflA and ORF_P of bacteriophage K5A exhibited 95% identity with the intergenic region between ORF_L and the endosialidase gene (Fig. 1A). This nucleotide sequence homology between noncoding regions of bacteriophage K5A and K1E indicates that these two bacteriophages are related. It was also observed that the first nine consecutive amino acids of the KflA and the ORF_L polypeptides were identical (Fig. 1B). The homology between these polypeptides could be extended to 41% identity (60% similarity) over the first 46 amino acids (Fig. 1B), although no homology was detected between the remaining 66 C-terminal amino acids of the ORF_L polypeptide and sequences in KflA. Translation of ORF_L has not been demonstrated, and the function of this ORF is unknown (22). However, the homology between the N termini of the ORF_L polypeptide and KflA suggests that the ORF_L may be a truncated remnant of kflA. Thus, it appears that bacteriophage K1E is a derivative of bacteriophage K5A in which the endosialidase gene was acquired by lateral transfer followed by the subsequent loss of the kflA gene. No homology was detected between the noncoding DNA 3′ of the K1E endosialidase gene and bacteriophage K5A sequence, and therefore, it is not known if the endosialidase gene was inserted between kflA and ORF_P or whether the gene represented by ORF_P was replaced by the endosialidase gene. In addition, it cannot be determined whether the endosialidase gene was acquired by K1E before or after loss of the kflA gene.

Expression and purification of a (His)₆-KflA fusion protein.

The KflA lyase was overexpressed in E. coli K-12 as a fusion protein containing an N-terminal (His)₆ tag. The entire kflA gene was amplified by PCR and cloned downstream of the araBAD promoter in the expression vector pBAD/His. The resulting plasmid was designated pLYA100. Transcription of kflA from the araBAD promoter is induced by the presence of l-arabinose in a dose-dependent manner (14, 15). Optimum expression of soluble KflA fusion protein was obtained by induction at an l-arabinose concentration of 0.02% (wt/vol) (Fig. 3).

FIG. 3.

Coomassie blue-stained gel showing overexpression of the (His)₆-KflA fusion protein in E. coli TOP10. Cells were induced with 0.02% (wt/vol) l-arabinose and grown for 2 h following induction. Lane 1, molecular weight markers (in thousands); lane 2, induced TOP10(pBAD/HisB); lane 3, induced TOP10(pLYA100); lane 4, purified (His)₆-KflA. The arrowhead indicates the purified KflA protein.

The (His)₆-KflA fusion protein was purified from 500 ml of induced culture using a combination of anion-exchange chromatography and immobilized metal (nickel) adsorption chromatography (IMAC). The majority of the KflA fusion protein was eluted from the ion-exchange column at an NaCl concentration of between 160 and 200 mM (results not shown). This material was pooled and subjected to IMAC. A protein, migrating at 66 kDa in SDS-PAGE, was eluted from the IMAC column at between 170 and 245 mM imidazole. This fraction was collected and shown to contain lyase activity (Table 1). The total activity obtained in the pooled IMAC fraction was greater than that obtained after the ion-exchange stage. It is possible that this increase in activity was due to removal of an inhibiting substance during the IMAC stage. A similar observation was found during the purification of a glycosyltransferase involved in chondroitin sulfate biosynthesis (39). After desalting and concentration, the KflA fusion protein was obtained at a 14-fold purification, giving a yield of 2.4 mg per 500 ml of original culture (Table 1). The fusion protein was judged homogeneous by SDS-PAGE (Fig. 3). A minor band was observed migrating slightly faster than the KflA fusion protein on SDS-PAGE (Fig. 3). However, this band became more prominent after prolonged storage at 4°C, and thus, the presence of this band was a result of degradation of the full-length fusion protein (data not shown).

TABLE 1.

Summary of purification of the (His)₆-KflA fusion protein from 500 ml of culture

Anion exchange	Total protein (mg)	Total activity (AU min−1)	Sp act (AU min−1 mg−1)	Yield	Purification
Crude lysate	122	869	7.12	100	1
Anion exchange	14	371	26.50	43	3.7
IMAC	3.2	389	121.56	45	17.1
Desalting	2.4	240	100.00	28	14.0

Analysis of (His)₆-KflA activity.

Cleavage of K5 polysaccharide with recombinant KflA was visualized by analyzing the products of the lyase reaction by PAGE. Untreated K5 polysaccharide preparations exhibit a wide range of molecular weights (Fig. 4, lane 1) as a result of varying levels of polymerization during biosynthesis. Incubation of K5 polysaccharide (800 μg/ml) with a dilution series of purified KflA fusion protein showed that at a high enzyme concentration (36 μg/ml) the polysaccharide was degraded into a single low-molecular-weight product (Fig. 4, lane 2). As the enzyme concentration was decreased from 9 to 2.25 μg/ml, the polysaccharide was not degraded to completion and products of successively higher molecular weight were accumulated (Fig. 4, lanes 3 to 6). Since it has been previously shown that incubation of bacteriophage K5A with K5 polysaccharide results in final reaction products of hexa-, oct-, and decasaccharides (12), the lowest-molecular-weight product observed following degradation with 36 μg of KflA per ml (Fig. 4, lane 2) is likely a hexasaccharide. Based on the β elimination reaction, the two sequentially higher-molecular-molecular weight bands (Fig. 4, lanes 4 to 7) represent octa- and decasaccharides, respectively. It has been suggested that cleavage of the K5 polysaccharide by the KflA lyase into oligosaccharides smaller than three repeating units may not be possible due to a minimum chain length required for substrate recognition and cleavage (12).

FIG. 4.

PAGE of K5 polysaccharide digested for 1 h with various concentrations of purified KflA lyase as described in Materials and Methods. Lane 1, undigested K5 polysaccharide; lane 2, 36 μg of KflA per ml; lane 3, 18 μg/ml; lane 4, 9 μg/ml; lane 5, 4.5 μg/ml; lane 6, 2.25 μg/ml; lane 7, 1.13 μg/ml; lane 8, 0.56 μg/ml; lane 9, 0.28 μg/ml; lane 10, 0.14 μg/ml.

Legoux et al. (16) demonstrated that the ElmA lyase produces final reaction products of 10 disaccharide units in length, considerably longer than those produced by KflA. The significance of this difference in the sizes of the reaction products is as yet unclear.

The purified KflA lyase fusion protein was active over a broad pH range (Fig. 5). Maximum activity was observed at pH 8.5, but activity diminished substantially as the pH was lowered below 6.0 or raised above 9.0. Although significant activity was observed at pH 6.75 and 7.25, the activity was decreased at pH 7.0 (Fig. 5). The reason for the low activity at pH 7.0 is unclear. However, the theoretical pI of native KflA is 7.09 [7.25 for (His)₆-KflA], and it is possible that a lack of charge at a pH approaching 7.0 may adversely affect KflA activity. The pH optimum of 8.5 differs markedly from an optimum of pH 5.5 found for the bacteriophage K1E endosialidase (37). Both bacteriophage K5A and K1E were isolated from sewage (7, 9), and presumably one of their natural habitats is the mammalian colon. The pH of the human colon ranges from 5.7 in the cecum to 6.7 in the rectum (4), and therefore, the KflA enzyme would be active in this environment. The KflA fusion protein had no apparent requirement for divalent cations, and activity was not significantly affected by changes in NaCl concentration. The enzyme retained most of its activity at an NaCl concentration as high as 1 M (results not shown).

FIG. 5.

Effect of pH on KflA lyase activity. Lyase activity was determined by monitoring the A₂₃₂ over 5 min. Reactions were performed using 1 μg of (His)₆-KflA fusion protein and 800 μg of K5 polysaccharide in each 1-ml reaction mixture. Reactions were performed in 25 mM morpholineethanesulfonic acid (MES) (pH 5.5 to 6.75)–25 mM HEPES (pH 7.0 to 8.0)–25 mM Tris-acetate (pH 8.5 to 9.0).

The K5 polysaccharide substrate utilized by the KflA lyase is identical to the precursor molecule of heparin and heparan sulfate (40). Synthesis of heparin and heparan sulfate involves N deacetylation and N sulfation of GlcNAc residues of [-4)-βGlcA-(1,4)-αGlcNAc-(1-]_n polymers followed by epimerization of GlcA residues to iduronic acid and O sulfation at various positions (17). These modifications are not stoichiometric, and therefore, heterogeneity exists due to the degree of epimerization and sulfation. Generally, heparin is more extensively modified than heparan sulfate. Heparin lyases produced by Flavobacterium heparinum are enzymes that degrade heparin and heparan sulfate (21). Three forms of heparin lyase have been purified and exhibit different specificities for heparin and heparan sulfate. Applications of these enzymes include structural studies of heparin-like polymers (18, 19), synthesis of low-molecular-weight therapeutic agents (20), and neutralization of heparin in blood. These enzymes do not have amino acid homology to KflA although they exhibit high pH optima similar to the bacteriophage enzyme (21). KflA recognizes heparan sulfate as a substrate, cleaving the polymer at regions devoid of iduronic acid and sulfation (K. Lidholt, personal communication). Thus, it is possible that the recombinant KflA lyase may be of use as a tool for structural studies of heparin and heparan sulfate and production of low-molecular-weight, highly modified heparin-like oligosaccharides.

In conclusion, the gene encoding a bacteriophage lyase (kflA) specific for the E. coli K5 capsular polysaccharide has been cloned and sequenced, and a His-tagged KflA fusion protein has been overexpressed and purified to homogeneity in an active form. The KflA lyase is not encoded on the E. coli chromosome, and kflA is distinct from the elmA polysaccharide lyase gene which is present on the chromosome of a subset of E. coli serotypes. Therefore, these two genes are distinct and either have been derived from separate progenitors or have diverged from a common progenitor for a considerable length of time. In addition, evidence has been provided which suggests that bacteriophage K5A is a progenitor to the E. coli K1 capsule-specific bacteriophage K1E.